Stories cluster related articles into a single narrative, linked to the underlying CVEs and affected products. 7,960 stories synthesized.
A 10-month Android malware campaign dubbed 'Premium Deception' used nearly 250 fake apps to silently enroll victims in premium SMS subscriptions via carrier billing, primarily targeting users in Southeast Asia and Eastern Europe.
Microsoft has open-sourced two new tools, RAMPART and Clarity, to help developers test and debug the security of AI agents during development.
Digital.ai's 2026 App Security Threat Report reveals that AI-assisted attacks against client-facing apps have risen from 55% in 2022 to 87% in 2026, with the iOS-Android security gap nearly closed and attack windows shrinking to hours.
1Password has partnered with OpenAI to introduce a just-in-time credential model for Codex, ensuring AI coding agents never persistently store secrets in prompts, code repos, or model context.
CISA added six Microsoft vulnerabilities to its Known Exploited Vulnerabilities catalog on May 20, 2026, including the Conficker worm flaw and two newly disclosed 2026 CVEs, all confirmed under active exploitation.
Anthropic quietly fixed a sandbox bypass in Claude Code that could be chained with prompt injection to exfiltrate data, raising concerns over disclosure practices.
Mozilla's Firefox 151.0 fixes a use-after-free vulnerability (CVE-2026-8953) that could allow sandbox escape, while introducing anti-fingerprinting measures and local network access prompts.
Akamai reports a surge in DDoS, API, and web attacks against global financial services firms in 2025, driven by AI-enabled botnets, hacktivists, and financially motivated cybercriminals.
A proof-of-concept exploit has been released for PinTheft, a Linux kernel privilege escalation vulnerability that primarily affects Arch Linux systems, allowing local attackers to gain root access.
AI-generated typosquatting domains are now used in supply-chain attacks, as demonstrated by the Trust Wallet Chrome extension compromise that drained $8.5M.
NLnet Labs released Unbound 1.26.0 on May 20, 2026, patching 11 vulnerabilities including a CVSS 10.0 cache-poisoning bug and a DNSSEC validator flaw enabling possible remote code execution.
Unit 42 has identified over 4,000 samples across three TamperedChef malware clusters that deliver stealthy payloads via trojanized productivity apps and malvertising.
A massive scareware campaign dubbed CypherLoc has targeted an estimated 2.8 million users since early 2026, using phishing emails and browser-locking techniques to push fake tech support scams.
A critical command injection vulnerability in ExifTool versions 13.49 and earlier on macOS allows attackers to execute arbitrary commands by embedding malicious metadata in image files.
ESET reveals that the China-linked APT group Webworm has expanded its cyber espionage operations beyond Asia, now targeting European government organizations with two new backdoors.
GitHub is investigating a claimed breach by threat actor TeamPCP, who listed approximately 4,000 internal repositories for sale on a cybercrime forum.
Ukraine's National Security and Defense Council reports Russia has dramatically expanded AI-powered malware and social engineering in cyber operations over the past year.
A zero-day vulnerability in Huawei enterprise routers caused a three-hour nationwide telecom outage in Luxembourg in July 2025, yet the flaw remains unpatched and undisclosed.
HUMAN's Satori team uncovered Trapdoor, an Android ad fraud and malvertising campaign involving 455 malicious apps that generated 659 million daily bid requests at its peak.
A months-long breach via a compromised third-party vendor has exposed highly sensitive data for at least 1.8 million patients and employees, including medical records, Social Security numbers, bank details, fingerprints, and palm prints.
Drupal is preparing an emergency patch for a 'highly critical' vulnerability, warning that exploit code could appear within hours or days of disclosure.
A critical authentication bypass vulnerability, CVE-2026-8598, has been disclosed in ZKTeco CCTV cameras, exposing account credentials and service information without authentication.
CISA disclosed four vulnerabilities in ScadaBR 1.2.0, including a CVSS 9.1 missing authentication bug that could allow unauthenticated remote code execution across critical infrastructure sectors.
CISA has issued an advisory for CVE-2025-3465, a path traversal vulnerability in ABB CoreSense HM and CoreSense M10 products that could allow unauthenticated local attackers to compromise systems and access sensitive data.